Team:Thrace/Engineering

In cooperation with our Dry Lab, our first option was to implement an electrochemical potentiostat, as in theory it achieves the best accuracy to cost ratio. The initial process of designing and electronically simulating a potentiostat model took many months, as we extensively studied the literature, past electrochemical potentiostat models such as SimpleStat etc. Due to Covid restrictions and global semiconductor shortages however, we had to adapt to a colorimetric model, despite all the research and simulation time already invested on a potentiostat model. Our team is highly skilled and even faced with the last minute seemingly inescapable odds we managed to proceed with a viable, even more economically viable alternative.

Using a photo standardization 3D printable cube that allows the use consumer photo sensors, cutting the cost for the final product considerable, and developing a color normalization algorithm of the photo taken to quantize the fluorescence of the test strip, enabling the use of an AI system in the future to classify the test cases, complementing the work of the medical personnel that is already hard at work at these difficult times

We started by discovering the physical restrictions of the colorimetric model, especially the non standard environmental lighting of each consumer, so the first iteration of the required solution was a folding dark paper mockup, that through testing we shaped accordingly to make the photo acquisition easy and as environment agnostic as possible. After finalizing the shape through several paper mockups and dozens of tests in a variety of lighting conditions, we arrived at the conclusion that a LED of known color temperature is needed as well as a mirror to prevent the light from the LED from blinding the camera sensor.

At the same time, the Dry lab was hard at work, modeling the optimal temperature used for target multiplication, the enzyme kinetics used by the Wet Lab, and the fluorescence to enzyme ratio of the final solution, using the Michael-Menten equations, while linearizing the results using the Lineweaver–Burk equation for easier representation. Using mathematical series and approximations we showed that fluorescence intensity is directly proportional to concentration, thus aiding th e further interpretation of the collected data.

The modeling process started by exploring all the popular mathematical options and implementations in enzyme kinetic modeling found in literature, testing the validity, runtime and consistency of each method, we developed an implementation on Matlab mathematical modeling software that is consistent, has high accuracy (>95%) and runs almost instantaneously. Stepping on the above milestones, we are ready to integrate a consumer device capable of detecting colon cancer with high accuracy while collecting anonymous data valuable to medical researchers and data analysts, further expanding the knowledge in the field, long after our device hits the market.

Wet Lab

We followed the engineering cycle regarding the primer (stem loop) that is used to reverse transcribe miR-766-8p. Firstly we created a modified stem loop primer from protocols [1,2] that had a melting temperature between 30 - 40 C in order for it to stay in its hairpin structure while the reverse transcription reaction takes place and to be linear while the recombinase polymerase amplification was in action. Unfortunately, after many tries, during our experiments, the reverse transcription and recombinase polymerase amplification of the cDNA did not work. Also the stem loop primer binds to only the last 6 bases at the 3’ end of the microRNA that we are trying to detect, thus there is room for improving the specificity of the design. So we tried to make a new stem loop primer that has a higher melting temperature (>50 C) than the first one in order to take advantage of a higher temperature reverse transcription reaction and increase the specificity for our intended microRNA target by binding to its 5’ and 3’ ends. The new stem loop primer has to have a hairpin structure with higher GC content and two tails at its 5’ and 3’ prime ends where the microRNA can bind. Actually, there is a design [1] that already exists but is made to work with PCR. Taking inspiration from that article and combining our knowledge about designing the previous stem loop primer that is compatible with RPA we designed the new primer and tested it using UNAfold DNA Folding Form [2] and this is the result:

References:

1. Androvic, P., Valihrach, L., Elling, J., Sjoback, R., & Kubista, M. (2017). Two-tailed RT-qPCR: a novel method for highly accurate miRNA quantification. Nucleic Acids Research, 45(15), e144-e144. https://doi.org/10.1093/nar/gkx588

2. DNA Folding Form. Unafold.org. (2021). Retrieved 21 October 2021, from http://www.unafold.org/mfold/applications/dna-folding-form.php.

Follow us:

SPONSORS